Ernst & Young Global Limited Becket House 1 Lambeth Palace Road London SE1 7EU Tel: +44 [0]20 7980 0000 Fax: +44 [0]20 7980 0275 www.ey.com International Accounting Standards Board 30 Cannon Street London EC4M 6XH 28 March 2013 Submitted electronically through the IASB website (www.ifrs.org) Dear IASB members Re: Invitation to comment Exposure Draft ED/2012/4 Classification and measurement: Limited amendments to IFRS 9 Proposed amendments to IFRS 9 (2010) The Global organisation of Ernst & Young is pleased to respond to the invitation of the IASB (the Board) to comment on the above Exposure Draft (ED or proposals). We support the Board s efforts to address certain questions raised by constituents on the application of the IFRS 9 classification and measurement model. However, we have a number of comments and concerns on the proposals. These are summarised below and are discussed more fully in appendix A to this letter. Convergence We have always been, and continue to be, strong supporters of the achievement of a single improved high quality global accounting standard on financial instruments, an objective that has been repeatedly called for by the G20 s leaders. We are concerned, therefore, that the joint efforts of the IASB and the FASB (the Boards) on the financial instruments projects have not been fully successful in achieving this outcome, in any of the three phases of the project, including classification and measurement. Accordingly, we encourage the Boards to continue to strive for a common solution for classification and measurement. In this regard, we note that the Basis for Conclusions cites the desire to reduce key differences from US GAAP as a significant basis for some of the proposals in this ED. However, as the Boards proposals are currently drafted, there are significant differences in the application guidance which could result in different outcomes for the same fact patterns. For example, differences in how the business model criteria are applied could lead to differences in classification and measurement conclusions. Not only would these differences, if unresolved, undermine the benefits of the attempted convergence as set out in the Basis for Conclusions, but financial statement users of each standard are likely to draw mistaken conclusions when trying to compare an entity applying IFRS 9 with one applying the FASB s proposed classification and measurement model. In our view, the quest for Ernst & Young Global Limited is a company limited by guarantee registered in England and Wales. No. 4328808
2 convergence should not only be at the level of the key principles, but also at that of the application guidance. Proposal to introduce FVOCI We appreciate that the fair value through other comprehensive income (FVOCI) measurement category is welcomed by many constituents as a preferred alternative to the profit or loss volatility or accounting mismatches that may arise if more instruments were to be classified at fair value through profit or loss (FVPL). However, we observe that such a solution raises a number of issues, as outlined below. First, the introduction of a new FVOCI category is not consistent with the Board s overarching objective of reducing the complexity of accounting for financial instruments. Furthermore, the proposal to recycle gains and losses differs from the treatment of investments in equity instruments, and so is potentially confusing and lacks technical coherence. While we recognise that this difference has arisen because of the difficulties in assessing impairment for equity instruments, we are concerned that the difference, in effect, creates a fourth category of financial instrument. Second, the objective to address potential accounting mismatches that would arise because of the interaction between the accounting for financial assets and the accounting for insurance contract liabilities has only been partially fulfilled. Insurers often include non plain-vanilla instruments, which do not qualify for FVOCI, in portfolios of assets that back their insurance liabilities. The recycling of the gains and losses on these assets will often not match the timing of gains and losses realised on the corresponding insurance liabilities. In addition, the Board s deliberations on the insurance project are still ongoing and, therefore, the intended matching may never materialise. Third, as already mentioned, the objective to converge with US GAAP has been only partly satisfied, given that key aspects of the application guidance on the boundaries of the measurement categories differ. Moreover, the divergence of the Boards impairment proposals means that convergence has not been achieved. We would, therefore, question the extent to which the proposals are consistent with the Board s original objectives. Ultimately, however, the Board must consider, in its re-deliberations, whether a new FVOCI category will best serve the needs of users while faithfully reflecting the business models used by preparers. Contractual cash flow characteristics assessment We support the Board s desire to improve the guidance for the characteristics of the asset assessment. However, we are concerned that the proposed modifications would not permit relatively common financial instruments issued in interest rate regulated environments to be recorded at amortised cost. In this regard, we support the IASB s intention to gather
3 further feedback on this subject and to work to find a solution for such instruments, where appropriate. Early application of own credit presentation requirements in IFRS 9 We welcome the proposal to allow early application of the own credit presentation requirements in IFRS 9. However, we recommend that the amendment is also made to IAS 39; otherwise it will only become available once IFRS 9 is completed and, in some jurisdictions when it is endorsed by the regulator. Other matters Appendix B to this letter includes other areas of interpretation or application that we recommend the IASB to consider in its re-deliberations or address as part of the minor amendment process. Should you wish to discuss the contents of this letter with us, please contact Robert McCracken on +44 (0) 20 7951 2026 or Tony Clifford on +44 (0) 20 7951 2250. Yours faithfully,
4 Appendix A: Answers to the specific questions Contractual cash flow characteristics assessment: a modified economic relationship between principal and consideration for the time value of money and credit risk Question 1: Do you agree that a financial asset with a modified economic relationship between principal and consideration for the time value of money and credit risk could be considered, for the purposes of IFRS 9, to contain cash flows that are solely payments of principal and interest? Do you agree that this should be the case if, and only if, the contractual cash flows could not be more than insignificantly different from the benchmark cash flows? If not, why and what would you propose instead? Question 2 Do you believe that this Exposure Draft proposes sufficient, operational application guidance on assessing a modified economic relationship? If not, why? What additional guidance would you propose and why? Question 3: Do you believe that this proposed amendment to IFRS 9 will achieve the IASB s objective of clarifying the application of the contractual cash flow characteristics assessment to financial assets that contain interest rate mismatch features? Will it result in more appropriate identification of financial assets with contractual cash flows that should be considered solely payments of principal and interest? If not, why and what would you propose instead? We support the proposal to allow plain vanilla financial assets with minor modification features to qualify for amortised cost measurement. We believe that the guidance provided is broadly operational, but it would be enhanced if the Board were to provide implementation examples showing how a detailed assessment would be performed. In particular, the Board should consider clarifying exactly how the assessment should be performed, including, for example, whether it is performed on a discounted or undiscounted basis. In this regard, we have prepared worked examples, which we have shared with the IASB Staff, illustrating that different interpretations of the assessment could potentially lead to different answers. To illustrate the appropriate benchmark instrument to which the cash flows of an instrument with a modified economic relationship between principal and interest should be compared, paragraph B4.1.9B provides the following example: if the financial asset under assessment contains a variable interest rate that is reset monthly to a three-month interest rate, the appropriate benchmark would be a financial asset with the identical contractual terms and the identical credit quality except that the variable interest rate is reset monthly to a monthly interest rate.
5 We suggest the IASB clarifies, in the Basis for Conclusions at a minimum, whether or not the benchmark instrument in this example can, alternatively, be a financial asset that is reset quarterly to a three-month interest rate. As indicated in paragraph 44 of the Basis for Conclusions, we note the concerns raised by constituents about some instruments whose cash flows are not solely principal and interest (in the narrower sense of IFRS 9) as a result of regulation. Although the interest rate may not reflect the economic notions of time value and credit risk, in a manner consistent with the concept in IFRS 9, such products are usually simple and do not introduce more complex risks for the holder such that amortised cost would not be an appropriate representation of the future cash flows. In this respect, we support the IASB s intention to gather further feedback. Depending on the outcome, we encourage the Board to consider amendments to IFRS 9 that would permit such instruments to be recorded at amortised cost, particularly when the interest rate structure as set by the government or central bank represents the pricing basis for domestic currency transactions in a specific jurisdiction. As drafted, the proposed benchmark instrument assessment would not apply to debt instruments with embedded derivatives, even when the effect of such embedded features is not significant. Such features may be included purely for jurisdictional tax reasons. In this regard, we note that the IASB Staff Paper 5A that was discussed at the February 2012 meeting states: "If the financial asset contains a feature (i.e. a "building block'') other than principal, compensation for the time value of money and the credit risk of the instrument, the instrument must be measured at FVPL. For example, that would be the case if interest payments are indexed to commodity prices or equity prices, even if the effect of such indexation is not expected to be significant. We do not understand why the benchmark instrument assessment should be limited to instruments with leverage or interest mismatch features on the grounds that their effect is insignificant when other modifications may be equally insignificant. We, therefore, suggest that the Board consider extending the benchmark instrument assessment to instruments with other insignificant modifications, possibly combined with a more freely available fair value option in order to avoid the need for entities to perform the assessment in all cases. Business model assessment: the fair value through other comprehensive income (FVOCI) measurement category for financial assets that contain contractual cash flows that are solely payments of principal and interest Question 4: Do you agree that financial assets that are held within a business model in which assets are managed both in order to collect contractual cash flows and for sale should be required to be measured at fair value through OCI (subject to the contractual cash flow characteristics assessment) such that:
6 (a) interest revenue, credit impairment and any gain or loss on derecognition are recognised in profit or loss in the same manner as for financial assets measured at amortised cost; and (b) all other gains and losses are recognised in OCI? If not, why? What do you propose instead and why? Although we acknowledge that the FVOCI measurement category may fulfill some of the objectives cited by the Board, and would be welcomed by many constituents, we question the extent to which its objectives have been achieved in light of the following concerns: (a) The FVOCI category would not completely address the concerns raised about the interaction between the accounting for insurance contract liabilities and the accounting for financial assets backing insurance contracts: (i) only plain vanilla debt instruments will qualify for this category, whereas a broad range of instruments is often used by insurers to back insurance liabilities; (ii) income statement mismatches would not be resolved if the financial assets are sold prior to their maturity (because of recycling); and (iii) under the impairment proposals, expected losses are recorded in the income statement when an instrument is first recognised, resulting in a day-one loss being recognised on the assets. Meanwhile, the Board s deliberations on the insurance project are still ongoing and, therefore, the intended matching may never materialise. (b) Whilst the introduction of FVOCI is intended, in part, to achieve convergence of IFRS 9 with the FASB s classification and measurement model, this objective is frustrated by the divergence in the two sets of impairment proposals as well as differences in the proposed application guidance regarding the scope of the three measurement categories. In particular, the application guidance relating to the fair value through net income (FVNI) business model in the FASB s proposals neither makes reference to portfolios managed on a fair value basis nor to held for trading portfolios. Rather, it states that only financial assets that do not meet the business model for either amortised cost or FVOCI classification would be measured at FVNI. The business model guidance for the amortised cost measurement category differs significantly under the two proposals. Moreover, the application guidance relating to the FVOCI category in the FASB proposals suggests that the business model assessment at initial recognition is made at the individual asset level. Such differences in the application guidance between the two proposals could potentially result in different classification outcomes. We urge the Boards to align their application guidance so that the classifications can be applied consistently with common boundaries between them. (c) Recycling of gains and losses upon the sale of financial assets, a key feature of this proposed FVOCI category, would result in reported performance being dependent on the timing of sales of financial assets. In addition, the mechanics of recycling, coupled with the application of the expected loss impairment model to the FVOCI measurement category, are not straightforward and will make the financial statements, particularly OCI, more difficult to understand.
7 (d) It appears hard to justify using two different FVOCI models in IFRS 9, one for equity instruments without recycling and another for debt instruments with recycling, especially when the Board is expected to address the issue of OCI as part of the Conceptual Framework project. 1 We acknowledge that there is a trade-off between the objective of simplification and meeting the other needs of different constituents. Ultimately, however, the Board must consider, in its re-deliberations, whether a new FVOCI category will best serve the needs of users while faithfully reflecting the business models used by preparers. If the Board decides not to proceed with the FVOCI category, it may wish to consider exploring other options. One possible approach is to introduce FVOCI as a targeted solution for insurers. Under this alternative, FVOCI would be introduced as an option to avoid accounting mismatches (but only as a consequential amendment to IFRS 9 and IFRS 4 once the insurance contracts project is finalised). This could possibly be combined with relaxing the hold to collect business model criteria for all entities so that more financial assets qualify for amortised cost measurement. 2 Business model assessment: the fair value through other comprehensive income measurement category for financial assets that contain contractual cash flows that are solely payments of principal and interest Question 5: Do you believe that the Exposure Draft proposes sufficient, operational application guidance on how to distinguish between the three business models, including determining whether the business model is to manage assets both to collect contractual cash flows and to sell? Do you agree with the guidance provided to describe those business models? If not, why? What additional guidance would you propose and why? 1 We recognise, however, that this difference has arisen because of the difficulties in assessing impairment for equity instruments. 2 The focus of this alternative would be on the reasons for sales, potentially combined with loosening the insignificant/infrequent sales criteria, but requiring clear disclosure of the reasons for sales. Such alternative would be accompanied by the existing requirement to disaggregate gains and losses on the face of the income statement, which would enable clear identification of gains and losses resulting from sales out of the amortised cost category, as well as a requirement to disclose fair values on the face of the balance sheet for instruments measured at amortised cost. This will facilitate users ability to determine gains and losses that would be realised upon sale of such instruments. Hence, the alternative approach would provide the same information as the proposed FVOCI measurement category. Supporters of the alternative approach acknowledge the difficulty in drawing a clear dividing line between amortised cost and fair value through profit or loss, but also argue that one dividing line is better than two. They also believe that a clear identification of gains and losses resulting from sales out of amortised cost category on the face of income statement would result in a strong natural pressure on management to limit sales out of this category.
8 If the IASB decides to introduce a FVOCI measurement category in the final standard, we recommend that it works jointly with the FASB to reach common criteria for the boundary between the FVOCI and the FVPL measurement categories. However, if the existing dividing line proposed in the ED between the FVOCI and fair value through profit or loss ( FVPL) measurement categories is retained, the clarity of the application guidance might be improved by removing the reference to held for trading portfolios. This is because held for trading portfolios are, by definition, a subset of portfolios managed on a fair value basis. We also suggest that the meaning of insignificant (level of sales) in paragraph B4.1.3 be further clarified. For example, should the insignificance assessment be made with reference to the average size of the portfolio over the average life of the portfolio, or should it be made, for example, by comparing the total gains and losses to be realised on derecognition to the total profit or loss of the reporting entity? Business model assessment: the fair value through other comprehensive income measurement category for financial assets that contain contractual cash flows that are solely payments of principal and interest Question 6: Do you agree that the existing fair value option in IFRS 9 should be extended to financial assets that would otherwise be mandatorily measured at fair value through OCI? If not, why and what would you propose instead? We support extending the FVPL option in IFRS 9 to financial assets that would otherwise be measured at FVOCI. However, as noted earlier, if the IASB decides to introduce a FVOCI measurement category in the final standard, we recommend that it aims to arrive at converged criteria with the FASB for the boundary between FVPL and FVOCI, and hence also for the scope of the FVPL option. Early application Question 7: Do you agree that an entity that chooses to early apply IFRS 9 after the completed version of IFRS 9 is issued should be required to apply the completed version of IFRS 9 (i.e. including all chapters)? If not, why? Do you believe that the proposed six- month period between the issuance of the completed version of IFRS 9 and when the prohibition on newly applying previous versions of IFRS 9 becomes effective is sufficient? If not, what would be an appropriate period and why? We support the proposal to require early adopters of IFRS 9 to apply the completed version of IFRS 9 once issued for the reasons specified in the Basis for Conclusions.
9 Early application Question 8: Do you agree that entities should be permitted to choose to early apply only the own credit provisions in IFRS 9 once the completed version of IFRS 9 is issued? If not, why and what do you propose instead? We welcome the proposal to allow early application of the own credit requirements in IFRS 9. However, we share the view of some constituents that the amendment should be made to IAS 39, otherwise it will only become available for many reporters (e.g., EU companies) once IFRS 9 is completed (and, in the EU, is endorsed). First-time adoption This Exposure Draft does not propose any specific changes to IFRS 1 First-time Adoption of International Financial Reporting Standards for first-time adopters of IFRS. However, to make sure that first-time adopters are given sufficient lead time to apply IFRS 9 and are not at a disadvantage in comparison to existing preparers, the IASB intends to consider the transition to IFRS 9 for first-time adopters when these proposals are re-deliberated. Question 9 Do you believe there are considerations unique to first-time adopters that the IASB should consider for the transition to IFRS 9? If so, what are those considerations? We agree with the Board s concern in paragraph BC113 of the ED that retrospectively applying some aspects of the completed version of IFRS 9 (especially impairment) would be impracticable due to the risk of applying hindsight, if those requirements were not actually applied during the reporting periods covered by the first IFRS financial statements. We also believe that the considerations that led the Board to give existing IFRS preparers relief from restating prior periods upon first applying IFRS 9 (as laid out in the Basis for Conclusions to the December 2011 amendments to IFRS 9) 3 are equally applicable to firsttime adopters. Accordingly, we support the Board s intention to reconsider the transition to IFRS 9 for firsttime adopters, once the re-deliberations of the proposed limited amendments to IFRS 9 and the impairment project progress sufficiently, to ensure that first-time adopters of IFRSs are given adequate lead time to apply IFRS 9 and are not at a disadvantage in comparison to existing preparers. 3 Paragraphs BC7.34A-M, Mandatory Effective Date and Transition Disclosures, Amendments to IFRS 9 and IFRS 7
10 Appendix B: Other areas of interpretation or application issues that we encourage the IASB to include within the scope of this project or address as part of the minor amendment process: 1. Business model assessment Loans held within a business intended for disposal IFRS 9 provides the following fact pattern as an example of a change in business model that would trigger a reclassification from amortised cost to FVPL: A financial services firm decides to shut down its retail mortgage business. That business no longer accepts new business and the financial services firm is actively marketing its mortgage loan portfolio for sale (emphasis added). This example assumes that the reporting entity has changed the business model of one of its business units from one that is hold to collect to one that is held for sale. The application question that we encourage the IASB to provide guidance on is whether the same conclusion would apply in a fact pattern where, for example, a bank makes a strategic decision to dispose of its auto finance business, which originates loans to collect their contractual cash flows. The bank intends to dispose of the entire business, including personnel, IT systems, intangibles and buildings, as a going concern and not merely a portfolio of loans. Assume this disposal does not yet fall within the scope of IFRS 5 Non-current Assets Held for Sale and Discontinued Operations. Despite the bank s intention to sell the business at some point in the future, the loans are still held within a business model whose objective is to hold them to collect their contractual cash flows. That objective continues regardless of whether or not the bank is able to sell the business. There is diversity of views on this issue. 2. Financial assets subsequently identified for sale We note that the FASB proposal contains specific guidance for financial instruments initially classified at amortised cost but that have been subsequently identified for sale for reasons other than changes in the business model (e.g., as a result of significant deterioration in the issuer s creditworthiness). Under this guidance, an entity would continue to classify the assets to be sold at amortised cost, but if the fair value of the financial asset identified for sale is below the amortised cost net of the allowance for credit losses (e.g., net carrying amount), an entity would record an impairment loss through net income. The impairment would be measured as the difference between the instrument s net carrying amount and fair value. The financial asset would also be identified on the face of the balance sheet as identified for sale. An entity would supplement that information with more detailed disclosures about why management departed from its held-for-collection business model. No similar guidance has been proposed in the IASB s ED. As a result, we encourage the Boards to work together to reduce differences on this point.
11 3. Guidance on instruments with non-recourse features A lender may originate or purchase a financial asset with non-recourse features that represent, in substance, either: (a) an investment in an underlying asset; or (b) a secured lending arrangement whereby the financial asset s cash flows do not vary with the performance of the collateral asset but rather give rise on specified dates to payments of principal and interest. There is limited guidance in IFRS 9 as to how to distinguish type (a) from type (b) instruments. We encourage the IASB to provide further application guidance on the following questions: 1. What is the definition of non-recourse? In particular, does it refer to the behaviour of the instrument only in default or bankruptcy, throughout the life of the instrument, or both? We note that clarifying the definition of non- recourse arrangements and the associated application guidance is necessary not only for financial assets like mortgages but also for single- or double-tranche structures that do not fall within the scope of the guidance on contractually linked instruments. We suggest the application guidance include illustrative examples of the criteria and considerations that need to be taken into account in assessing the substance of the non-recourse arrangement. In this regard, we note that the intention to control the collateral notion used in paragraph 48(b) of the Basis for Conclusions of the ED provides one useful consideration. 2. How does one make the assessment required by IFRS 9.B4.1.16 and 17? Take the example of a sub-prime mortgage loan, granted to a borrower of low creditworthiness for a high proportion (say 95%) of the value of the secured property. In the event of default by the borrower, the lender s only recourse is to the property asset. Should an assessment be made of the capacity of the borrower to service the loan, since the less likely it is that the borrower can service the loan, the more likely it is that the cash flows received by the lender will come from sale of the property asset? Should the level of collateralisation be considered? We encourage the Board to provide more application guidance clarifying the features that would render an instrument with a non-recourse provision to be an investment in the underlying asset rather than merely a secured lending arrangement. 3. Should structures that achieve the same substance as non-recourse loans, e.g., through the use of subsidiaries that borrow money and have limited assets, be treated the same way? 4. Classification of instruments prepayable at an amount different from the funded amount Whilst IFRS 9 does not currently define the term principal amount, the Basis for Conclusions states that cash flows that are interest always have a close relation to the amount advanced to the debtor (the funded amount). Furthermore, the Staff paper 5A
12 presented at the IASB and FASB joint meeting in February 2012 notes that principal is understood as the amount transferred by the holder on initial recognition. The strict application of this description would potentially result in loans advanced or debt instruments purchased at a premium or a discount and prepayable at par plus accrued interest (without penalty) failing the contractual cash flow characteristics test, irrespective of the size of the premium or discount. We encourage the IASB to provide further guidance on the application of the contractual cash flow characteristics test to such instruments. Further, we encourage the IASB to include a definition of principal amount in IFRS 9. 5. Instruments with a mandatory (automatic) prepayment provision It is common for loan agreements to include a mandatory repayment acceleration provision when the borrower breaches any term of the loan agreement. Such accelerated repayment provision is not conditional on the lender demanding it but happens automatically. The application issue that we encourage the IASB to clarify, as part of a future amendment, is whether it intended such mandatory prepayment provisions be scoped out of the exceptions in paragraph IFRS 9.B4.1.10 and B4.1.12, which refer only to prepayment options permitting the holder (i.e., the creditor) to put a debt instrument back to the issuer. 6. Contractually linked instruments indirect non-financial exposure Paragraph BC48 (b) states (...) the possibility that the pool may contain the collateral in the future should be disregarded unless the instrument was acquired with the intention of controlling the collateral. This is consistent with the manner in which collateral underlying financial assets is considered more generally for classification purposes, i.e., that a financial asset that is collateralised can still have payments that consist solely of principal and interest. (Emphasis added) We have the following observations on this paragraph: The underlined exception is not included within the body of the standard or its application guidance. We suggest the IASB include this exception in paragraph B4.1.26. The second half of the paragraph clarifies that the treatment of collateral underlying contractually linked instruments (e.g., collateral underlying a CDO tranche) is similar to the treatment of collateral underlying financial assets (e.g., collateral underlying a mortgage). However, we note that when the issuer of a financial asset defaults and the holder of the financial asset forecloses on the underlying collateral (e.g., the property in the case of a mortgage), the holder would recognise a non-financial
13 exposure on its statement of financial position. In contrast, even if a sizeable number of the assets underlying a contractually linked instrument (e.g., the properties underlying a portfolio of mortgages that serves as the collateral for a tranche) are repossessed, such non-financial exposure would not be as visible. Hence, we suggest that the IASB consider whether it is appropriate to require additional disclosures or a different accounting treatment for indirect non financial exposures when such exposure exceeds a certain tipping point. *************************